The geography of R&D: tobit analysis and a Bayesian approach to mapping R&D activities in the Netherlands

Regions vary strongly according to the participation of firms in R&D activity. By linking data on R&D activity at the firm level with GIS-based data on economic and other location features of regions, we are able to investigate the impact of local factors on R&D involvement for various types of firms. The relative importance of local factors as determinants of the R&D involvement of firms is estimated by means of a tobit model. Rather strong differences are found between zones in the same urban region. For example, modern manufacturing firms located in the centre of large cities have relatively low levels of R&D, and the opposite holds for rings of zones at certain distances from the cities. Bayesian methods are used for map presentations of the survey data.